Reliability and safety analysis methodology for identification of drivers’ erroneous actions

Owing to significantly individual differences in everyday driving behavior, it is quite difficult to assess the relative importance of driver errors compared with vehicle faults or road environment anomalies. This paper briefly presents several basic concepts for analysis of driving dependability including driving errors, driving reliability, driver recovery from erroneous actions, and key factors that shape driving behavior. This presentation is followed by construction of a shaping architecture for driving behavior that consists of a perception stage, a decision-making stage, an execution stage and correlativity among stages, in addition to internal feedback from complex traffic states. The causation classification of driving errors is then discussed in the context of three elemental types: perception error, decision-making error and execution error. The emphasis of this paper is on how to quantify driving dependability in order to identify various erroneous driver actions during traffic accidents. Specifically, this paper proposes a methodology to measure the probability of driving errors by considering the driver recovery from erroneous actions. The purpose of model-based driving dependability analysis is to quantitatively and qualitatively analyze the relationship between driving errors and traffic accidents causations.     

Design factor optimization of 3D flash lidar sensor based on geometrical model for automated vehicle and advanced driver assistance system applications

Sensor technologies have been innovated and enhanced rapidly for highly automated vehicle and advanced driver assistance systems (ADAS) in automotive industry; however, in order to adopt sensors into mass production vehicle in near future, various requirements should be satisfied such as cost, durability, and maintainance without any loss of overall performance of the sensors. In this sense, a 3D flash lidar is one of primising range sensors because of no moving parts, compact package, and precise measure for distance by using a laser. In spite of the several advantages, the 3D flash lidar is not commonly used in automotive industry because it is quite expensive for adoption and it is manufactured with only general purpose currently; therefore, the cost reduction and optimal design to satisfy various purposes of ADAS or autonomous driving should be accomplished. In this paper, we propose a novel approach for design factor optimization of the 3D flash lidar based on a geometrical model by using structural similarity between the 3D flash lidar and 2D digital camera. In particular, focal length and area of a receiver (focal plane array and read-out integrated circuit) which directly affect on sensor performance (field of view and maximum detection range) are optimized as the design factors. From the optimization results in simulation, we show that optimal design factors according to various purposes required in ADAS could be easily determined and the sensor performances could be evaluated before manufacturing. It will reduce temporal and economic burdens for design and manufacturing in development process.     

Experimental study on the characteristics of nano-particle emissions from a heavy-duty diesel engine using a urea-SCR system

Particulate matter in diesel engine exhaust, particularly nano-particles, can cause serious human health problems including diseases such as lung cancer. Because diesel nano-particle issues are of global concern, regulations on particulate matter emissions specify that not only the weight of particulate matter emitted but also the concentration of nanoparticles must be controlled. This study aimed to determine the effects on nano-particle and PM emissions from a diesel engine when applying a urea-SCR system for NO_x reduction. We found that PM weight increases by approximately 90% when urea is injected in ND-13 mode over the emission without urea injection. Additionally, PM weight increases as the NH₃/NO_x mole ratio is increased at 250 °C. In SEM scans of the collected PM, spherical particles were observed during urea injection, with sizes of approximately 200 nm to 1 μm. This study was designed to determine the conditions under which nano-particles and PM are formed in a urea-SCR system and to relate these conditions to particle size and shape via a quantitative analysis in ND-13 mode.     

A comparative study on the power characteristics and control strategies for plug-in hybrid electric vehicles

A comparative study was performed on two types of plug-in hybrid electric vehicles (PHEVs): the GM Volt and the Toyota Prius Plug-in Hybrid. First, the powertrain models of the two vehicles were derived. Based on the dynamic models, a detailed component control algorithm was developed for each PHEV. Specifically, a control algorithm was proposed for motor generator 1 (MG1) and MG2 to achieve optimal engine operation. Additionally, an energy management strategy for selecting the operation mode was developed from the viewpoint of fuel economy, battery state of charge and vehicle velocity. Using the dynamic model of the control algorithm for each PHEV, simulations were performed, and the simulation results were verified by comparing them with those obtained using the Powertrain System Analysis Toolkit simulator for the plug-in Prius. Based on the simulation results, a comparative study was performed, and it was found that the role and capacity of MG1 and MG2 and the mode selection algorithm must be determined depending on the configuration of the PHEV.     

Numerical investigation of a DI diesel spray flame using a detailed chemical reaction mechanism

Emission standards have grown increasingly stricter, consequently triggering greater interest in issues surrounding environmental pollution. In particular, soot and NOx released from DI diesel vehicles is considered to be the main source of air pollution in urban environments. However, the mechanics of fuel spray formation and the influence of the operating parameters on the resulting spray flame are not yet fully understood. In this study, the original KIVA code was modified to incorporate a detailed chemical reaction mechanism involving various species and multiple reaction steps to better understand the spray characteristics. n-Heptane, C₇H₁₆, was used as the representative fuel for diesel fuel, and the reaction mechanism for this fuel was composed of 66 species and 274 elementary reaction steps. The accuracy of the predicted results was demonstrated primarily by a comparison with experimental results. The numerical prediction of a specific operating condition for the parametric investigation correlates well with the experimental results.